BS 6364 vs ISO 28921-1: what a cryogenic test proves

BS 6364 is a British Standard for valves in cryogenic service. It sets design requirements and a low-temperature type test, and buyers name it in a purchase order when they want a valve design qualified at cryogenic temperature.

ISO 28921-1 is the international standard for isolating valves in low-temperature applications. Part 1 covers design, manufacturing and production testing, and it is written around the valves supplied in an order rather than only around a representative design.

Neither document is a certificate in itself. They are specifications: a report states which standard, edition and test conditions were used, and that statement is what you compare with your purchase order.

The table summarises how the two documents are normally used in a valve order. Test temperature, tracer gas and acceptance limits are defined by the edition named in the purchase order, so read that edition before you compare reports.

A cryogenic test is run at the valve’s minimum design temperature, or at the test temperature agreed in the order. Liquid nitrogen at about −196 °C is the usual bath for LNG and liquid-nitrogen service, but a valve rated to a warmer minimum, such as −46 °C, is not automatically tested at −196 °C.

The report must record the temperature actually reached, how it was measured and how long the valve was held at temperature. A report that says only “cryogenic tested” cannot be compared with a specification.

Cryogenic leak tests use a tracer gas and a mass spectrometer rather than a water bath and a bubble count. Helium is the usual tracer because its small molecule passes small leaks, and the spectrometer measures the rate directly.

A hydrostatic shell test in water and a helium leak test answer different questions. The shell test proves strength at ambient temperature; the helium test shows how tight the seat and body seals are when the valve is cold.

The result of a cryogenic leak test is a rate, measured with a mass spectrometer and compared with a limit in the standard named in the order. Rates are reported in units such as mbar·l/s or cm³/s, and the record should state the unit, the gas, the test pressure and the acceptance limit.

Do not compare that rate with the bubble count or “no visible leakage” wording of an ambient seat test. They are different tests with different limits, and a lower number in one does not prove the other.

Ask for the test record for the serial number in the order, not only a summary certificate. It should identify the instrument and its calibration, the temperature, the pressure, the measured rate and the pass limit.

A type test qualifies a design. One representative valve, or a range agreed in the test programme, is tested at low temperature, and the report then covers the sizes, classes and seat designs it names. It says nothing about the individual valve in your order.

Production testing is applied to the valves you buy: shell and seat tests at ambient temperature, a functional or operating-torque check where the standard requires it, and any low-temperature test the order adds. Each valve gets its own record with its serial number.

An order that needs both should say so: a BS 6364 type-test report for the design and production test records for each valve. One does not replace the other.